Potential of Aerial Image-Based Monoscopic and Multiview Single-Tree Forest Inventory: A Simulation Approach

Author： Korpela Ilkka Sakari Tokola Timo Erkki

Publisher： Society of American Foresters

ISSN： 0015-749X

Source： Forest Science, Vol.52, Iss.2, 2006-04, pp. : 136-147

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Two aerial image-based estimation chains were compared for their potential in the assessment of timber resources. The two chains consisted of two- and three-dimensional image measurements followed by indirect model estimation. The effects of various errors at both the single-tree and stand levels were examined with Monte Carlo simulation. The three-dimensional multiview estimation chain with height measurements of individual trees proved less prone to random and systematic errors than the monoscopic chain. An upper limit of accuracy at the single-tree level is determined mainly by the allometric correlations. Volume estimates of single trees showed estimation accuracies (root-mean-square error, RMSE) of ∼15–20% and 35–40% in the multiview and monoscopic chains, respectively. This is lower than the accuracy achievable in the field, which is ∼6%. Image-based single-tree forest inventory is prone to systematic measurement and model errors, and in a simulation for the total effect of error sources, upper estimation accuracies (RMSE) of ∼15% and 30% for stand volume were computed using full mapping of the stand with the multiview and monoscopic estimation chains, respectively.